Jae-Young Suh

We don’t have enough information about this author to calculate their statistics. If you think this is an error let us know.
Learn More
In the experiment, we evaluate the audio-only and the selected lip feature based visual-only speech recognition performances separately. For audio-only speech recognition, we use HTK3.2 [4], and adopt normalized log likelihood ratio (LLR) scores as the N-best confidence scores. For the visual part, we build a back propagation neural network by using SNNS(More)
  • 1